-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/8026/#review14097
-----------------------------------------------------------


Since this will be getting included in the upcoming release, I've tested this 
on several platforms.

Works for:
  OSX 10.7
  Ubuntu 12.04 (64 bit) * (make check broken from code off trunk)
  Ubuntu 10.04 (64 bit) ** (make check broken from code off trunk)
  CentOS 5.5

* Although this had a test failure on code in trunk (unrelated to this change):
[ RUN      ] ProcTest.Cpus
../../src/tests/proc_tests.cpp:57: Failure
Expected: (1u) <= (cpus.get().size()), actual: 1 vs 0
[  FAILED  ] ProcTest.Cpus (3 ms)

Contents of /proc/cpuinfo for my 12.04 VM
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 58
model name      : Intel(R) Core(TM) i7-3520M CPU @ 2.90GHz
stepping        : 9
microcode       : 0x15
cpu MHz         : 2891.326
cache size      : 4096 KB
fpu             : yes
fpu_exception   : yes
cpuid level     : 13
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov 
pat pse36 clflush dts mmx fxsr sse sse2 ss syscall nx rdtscp lm constant_tsc up 
arch_perfmon pebs bts nopl xtopology tsc_reliable nonstop_tsc aperfmperf pni 
pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt aes xsave avx f16c rdrand 
hypervisor lahf_lm ida arat epb xsaveopt pln pts dtherm fsgsbase smep
bogomips        : 5782.65
clflush size    : 64
cache_alignment : 64
address sizes   : 40 bits physical, 48 bits virtual
power management:

It doesn't contain the 'physical id' and 'core id' that we look for in the 
cpuinfo parsing. This code won't be in the release though right?



** Had to install libz via libcurl4-openssl-dev, configured correctly but the 
build is broken again due to the proc code (unrelated to this change):

../../src/linux/proc.hpp: In member function ‘virtual void 
CgroupsCpusetTest_IntegerAllocations_Test::TestBody()’:
../../src/linux/proc.hpp:73: error: dereferencing pointer ‘<anonymous>’ does 
break strict-aliasing rules
../../src/linux/proc.hpp:78: error: dereferencing pointer ‘<anonymous>’ does 
break strict-aliasing rules
../../src/linux/proc.hpp:68: error: dereferencing pointer ‘<anonymous>’ does 
break strict-aliasing rules

And the code:

55 struct CPU
56 {
57   CPU(unsigned int _id, unsigned int _core, unsigned int _socket)
58     : id(_id), core(_core), socket(_socket) {}
59
60   bool operator == (const CPU& that) const
61   {
62     return (id == that.id) && (core == that.core) && (socket == that.socket);
63   }
64 
65   bool operator < (const CPU& that) const
66   {
67     // Sort by (socket, core, id).
68     if (socket != that.socket) {
69       return socket < that.socket;
70     }
71 
72     // On the same socket.
73     if (core != that.core) {
74       return core < that.core;
75     }
76 
77     // On the same core.
78     return id < that.id;
79  }
80 
81   // These are non-const because we need the default assignment operator.
82   unsigned int id; // "processor"
83   unsigned int core; // "core id"
84   unsigned int socket; // "physical id"
85 };

Either way, this code will not be in the release you're cutting right ben?

- Ben Mahler


On Dec. 6, 2012, 6:22 p.m., Ben Mahler wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/8026/
> -----------------------------------------------------------
> 
> (Updated Dec. 6, 2012, 6:22 p.m.)
> 
> 
> Review request for mesos, Benjamin Hindman and Vinod Kone.
> 
> 
> Description
> -------
> 
> Now, configure.ac takes care of:
>   -Detecting python >= 2.6 automatically (Added python.m4 from automake 1.10 
> tweaked to include python2.6)
>   -Linking with libunwind when present.
>   -Including some special java include paths from the old configure.macosx
> 
> python.m4 is only necessary if we intend to support automake 1.10.
> automake 1.11 is aware of python 2.7, whereas 1.10 is only aware of python 
> 2.5.
> 
> 
> This addresses bug MESOS-285.
>     https://issues.apache.org/jira/browse/MESOS-285
> 
> 
> Diffs
> -----
> 
>   Makefile.am 5fae5c3a35960bd52619841b438f77e1b9868148 
>   configure.ac e6688d7e1e9f09d2d3a2c1c18e920371d47208e0 
>   configure.amazon-linux-64 716153358144395f78bcdddb0c8559d9691c27e9 
>   configure.centos-5.4-64 0a33ff0961ea6a9893a4e4aa738f53c48297bafe 
>   configure.macosx e9a80779367411f1a045143c2ed99f2d04c816f1 
>   configure.ubuntu-lucid-64 496c16201026ea7693a1674667fbf29ed2b3c570 
>   configure.ubuntu-natty-64 f940128ecfa4e6cdf9e22f9ebf48486ab5e01205 
>   m4/python.m4 PRE-CREATION 
>   m4/python.m4.patch PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/8026/diff/
> 
> 
> Testing
> -------
> 
> Tested on OSX 10.7 and CentOS 5.
> 
> 
> Thanks,
> 
> Ben Mahler
> 
>

Reply via email to